LOD, or Level of Development, is the system for classifying the amount of geometric detail within a model as set forth by the AIA’s E202 documents. James Vandezande has a great post on it from (believe it or not) 2008. The post does a great job at describing the different levels included, 100 to 500, and gives good reference to their intended meaning. Having used this systems on several projects now, I wanted to offer some feedback.
A few years back, when beginning to work on the Louisiana Museum and Sports Hall of Fame, I was inundated with emails and calls from just about every trade trying to figure out what constituted an acceptable deliverable. As the BIM Manager for the project, we were in charge of defining the requirements for a model deliverable that would be efficient, but also meeting the spec requirement for an ‘algorithmic coordination process’.
That project had another requirement that had a big impact on this conversation. The more than 1,000 unique cast stone panels had a surface integrity performance criteria that could only be achieved through CNC manufacturing. The fabricator, then, would be forced to model the panels and then use these models for mold making. This created a precedent. One of the more difficult elements in the design didn’t even have a choice but to create fabrication-level models. Did this requirement extend to the rest of the trades?
What if a trade wasn’t going to use the model for fabrication? What if they didn’t have the capacity to do so in their shops?
What became instantly obvious was that there was an ocean of space for interpretation in the LOD system. The specs called for an LOD400 model to be developed in the service of coordination. An LOD400 model is one that contains ‘shop drawing’ or ‘fabrication’ level of information. There was one big problem with this definition on this project, however. The lack of an explicit requirement on the model’s intended use left a gaping hole in the spec for a participant to interpret as they saw fit. One company’s view of what makes a good shop drawing versus another can be very different.
The lesson then was to not just ask for a deliverable, but to also ask for that deliverable to meet a certain level of performance. A model used for fabrication will inevitably make for a model that is accurate for coordination purposes. The opposite, however, is not always the case.
In this project, the hook was to also ask that all information presented in shops (the legal document) also be present in the model. This way, the architect could use the model as a real reference in their shop drawing review process. The importance of this requirement was that it ties their legal deliverable with the BIM process. This made BIM central to their every day project management knowing the architect could reject a submittal by virtue of their model quality or completeness.
This move though would not work on many projects and is hard to enforce. A better way to define model performance is to actually tie the deliverable to a good performance specification. That performance specification would include a reference to LOD. It would also include a responsibility matrix (the MET in the E202), and an intended use matrix. These three elements together make for a much better definition for model deliverables.
A good BIM execution plan (BEP) should include all these documents and a framework for their implementation.